제 2유형 [회귀 Regression] 자전거 수요 예측

이대로만 하면 2유형 40점 만점!

https://www.kaggle.com/code/agileteam/t2-6-bike-regressor

T2-6. Bike-Regressor

 

import pandas as pd

train = pd.read_csv("/kaggle/input/bike-sharing-demand/train.csv")
test = pd.read_csv("/kaggle/input/bike-sharing-demand/test.csv")

#target: count컬럼

#EDA
# print(train.info())
# print(train.head())
# print(test.info())
# print(test.head())

'''
train과 test의 컬럼 맞추기 ... 
없애야 할 독립변수: casual, registered
종속변수: train['count']

결측치 없음
datetime pd.to_datetime해주어야함
레이블인코딩 필요없음

이상치 확인 .. 없음!

# print(train['season'].unique())
# print(train['holiday'].unique())
# print(train['workingday'].unique())
# print(train['weather'].unique())
'''
# #데이터 전처리 - 결측치 처리 : 생략!
# for column in train.columns:
#     mode_val = train[column].mode()[0]
#     train[column].fillna(mode_val, inplace = True)
    
# for column in test.columns:
#     mode_val = test[column].mode()[0]
#     test[column].fillna(mode_val, inplace = True)
train['datetime'] = pd.to_datetime(train['datetime'])
test['datetime'] = pd.to_datetime(train['datetime'])

train['year'] = train['datetime'].dt.year
train['month'] = train['datetime'].dt.month
train['day'] = train['datetime'].dt.day
train['dayofweek'] = train['datetime'].dt.dayofweek
train['weekend'] = train['dayofweek'].apply(lambda x: 1 if x>=5 else 0)

test['year'] = test['datetime'].dt.year
test['month'] = test['datetime'].dt.month
test['day'] = test['datetime'].dt.day
test['dayofweek'] = test['datetime'].dt.dayofweek
test['weekend'] = test['dayofweek'].apply(lambda x: 1 if x>=5 else 0)



#데이터 전처리 - 레이블인코딩 
# from sklearn.preprocessing import LabelEncoder

# le = LabelEncoder()
# for column in train.select_dtypes(include = ['object']).columns:
#     train[column] = le.fit_transform(train[column])

# for column in test.select_dtypes(include = ['object']).columns:
#     test[column] = le.fit_transform(test[column])

'''
datetime에서 year, month, dayofweek만 사용!
'''


print(train.head())
print(test.info())

#데이터 분할
from sklearn.model_selection import train_test_split

X = train.drop(columns = ['datetime','casual', 'registered', 'count'])
y = train['count']

x_train, x_test, Y_train, Y_test = train_test_split(X, y, test_size = 0.2, random_state = 11)

#모델링, 학습, 예측
from sklearn.ensemble import RandomForestRegressor

rfr = RandomForestRegressor(n_estimators = 20, max_depth = 20, random_state = 11)
rfr.fit(x_train, Y_train)
pred1 = rfr.predict(x_test)

# 예측 성능 평가
from sklearn.metrics import mean_squared_error, r2_score
mse = mean_squared_error(Y_test, pred1)
rmse = mse**0.5
print(rmse)

r2 = r2_score(Y_test, pred1)
print(r2)

# 테스트데이터로 예측
test_X = test.drop(columns = ['datetime'])
pred2 = rfr.predict(test_X)

#결과 데이터 생성 및 확인
pd.DataFrame({'datetime': test['datetime'], 'pred': pred2}).to_csv('result.csv', index= False)
result = pd.read_csv('result.csv')

print(result)

 

---

데이터 전처리 반복문으로 한번에 !

 

1. 결측치 처리

for column in df.columns:
	mode_value = df[column].mode()[0]
	df[column].fillna(mode_value, inplace = True)

 

2. 레이블 인코딩

from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()

for column in df.select_dtypes(include = [‘object’]).columns:
	df[column] = le.fit_transform(df[column])